Vous êtes sur la page 1sur 5

International Epidemiological Association 2002 Printed in Great Britain International Journal of Epidemiology 2002;31:12351239

FETAL ORIGINS

Fetal origins of adult disease: strength of


effects and biological basis
DJP Barker,a JG Eriksson,b T Forsnb,c and C Osmonda

Background Low birthweight has been consistently shown to be associated with coronary
heart disease (CHD) and its biological risk factors. The effects of low birthweight
are increased by slow infant growth and rapid weight gain in childhood. To
quantify the importance of developmental processes in the genesis of CHD it is
necessary to establish the impact of fetal, infant and childhood growth on major
pathological events in later lifedeath, hospital treatment and the need for
medication.
Methods Longitudinal study of 13 517 men and women who were born in Helsinki
University Hospital during 19241944, whose body sizes at birth and during
childhood were recorded, and in whom deaths, hospital admissions, and
prescription of medication for chronic disease are documented.
Results The combination of small size at birth and during infancy, followed by accelerated
weight gain from age 3 to 11 years, predicts large differences in the cumulative
incidence of CHD, type 2 diabetes and hypertension.
Conclusions Coronary heart disease and type 2 diabetes may originate through two
widespread biological phenomenadevelopmental plasticity and compensatory
growth.
Keywords Fetal growth, childhood growth, type 2 diabetes, hypertension, coronary heart
disease
Accepted 11 June 2002

Associations between small body size at birth and during nephron numbers at birth, a known correlate of low birth-
infancy and later cardiovascular disease and its biological risk weight.8 In order to quantify the importance of developmental
factors have been found consistently.14 For different risk processes in the genesis of cardiovascular disease and type 2
factors the magnitude of the effects varies. For example, while diabetes we need to establish the impact of fetal and infant
birthweight is associated with large variations in indices of growth on major pathological events in later life rather than risk
insulin resistance,5,6 it is associated with small variations in factors.
blood pressure.7 A possible explanation for this is that, following Recent studies have shown that the effect of small body size
an intra-uterine lesion regulatory mechanisms may maintain at birth on later cardiovascular disease and type 2 diabetes is
homeostasis for many years until further damage, due to age, modified by the path of childhood growth.912 In particular,
obesity or other influences, initiates a self-perpetuating cycle of rapid childhood weight gain increases the risk of disease
progressive functional loss.8 Brenner has proposed such a model associated with small body size at birth and during infancy.
for the development of hypertension following reduced Examination of the risk of disease attributable to early
development therefore requires data on fetal, infant and
childhood growth. The Hertfordshire studies showed, for the
a Medical Research Council Environmental Epidemiology Unit, Southampton first time, that people who had low birthweight and low weight
SO16 6YD, UK. at age one year were at increased risk of developing coronary
b National Public Health Institute, Department of Epidemiology and Health
heart disease (CHD) and type 2 diabetes;1,13 but the data did
Promotion, Diabetes and Genetic Epidemiology Unit, Mannerheimintie
not include measurements of height or weight beyond 1 year
166, Helsinki, Finland.
c University of Helsinki, Department of Public Health, 00014 Helsinki, Finland. of age. We have examined the combined effects of fetal, infant
and childhood growth in the Helsinki cohort and present data
Correspondence: Professor DJP Barker, MRC Environmental Epidemiology
Unit, University of Southampton, Southampton General Hospital, on the magnitude of the effects together with their possible
Southampton SO16 6YD, UK. E-mail: djpb@mrc.soton.ac.uk biological interpretation.

1235
1236 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Methods predicted CHD more strongly than birthweight,9 while among


women birth length was the strongest predictor.10 We therefore
The Helsinki cohort comprises a group of men and women, analysed men and women separately using these two different
born 19241933, for whom there are data on size at birth and measures of body size at birth.
serial measurements of height and weight through school years;
together with another, younger group of men and women born Results
19341944, for whom there are serial measurements of height
Type 2 diabetes and hypertension
and weight from birth.12 The total cohort comprises 15 846
individuals: we present here data for 13 517 men and women, Table 1 is based on 698 patients being treated for type 2 diabetes
for whom there are measurements of body size both at birth and 2997 patients being treated for hypertension in people born
and at age 11 years. In Finland there are national registers of 19241944. It shows odds ratios according to birthweight and
hospital discharges by cause, and registers of people receiving fourths of BMI at age 11 years. The risks for each disease fell
medication for chronic illness, including type 2 diabetes and with increasing birthweight and rose with increasing BMI.
hypertension, in addition to registers of deaths by cause.912 The odds ratio for type 2 diabetes was 0.67 (95% CI: 0.580.79)
The registers of people receiving medication are maintained for each kilogram increase in birthweight and 1.18 (95%
because the costs of medication are reimbursed to the patients. CI: 1.131.23) for each kg/m2 increase in BMI at age 11 years.
This is subject to the approval of a physician who reviews each The corresponding figures for hypertension were 0.77 (95% CI:
case history. The diagnosis of type 2 diabetes is based on World 0.710.84) and 1.07 (95% CI: 1.041.09).
Health Organization criteria.14 Of the estimated 150 000 Table 2 is based on 227 patients being treated for type 2
patients with diagnosed diabetes in Finland 113 000 (75%) are diabetes and 1036 patients being treated for hypertension in
treated with medication.15 It is therefore possible to relate fetal people born 19341944. Subjects are divided into six groups
and childhood growth to type 2 diabetes and hypertension according to thirds of birthweight and whether their standard
defined by prescription of medication,11,12 and to CHD defined deviation score for BMI decreased or increased between age 3
by hospital admission or death.9 We identified all people and 11 years. For both diseases there were independent effects
receiving medication for type 2 diabetes and hypertension at of birthweight and change in BMI score. In the group with
any time from 1964 to 1997 and all hospital discharges or the highest birthweight and a subsequent decrease in BMI
deaths from CHD during 1971 to 1997. score, the cumulative incidence of type 2 diabetes was 1.5%
We calculated odds ratios for type 2 diabetes and (95% CI: 0.92.4%). This was less than half the incidence in
hypertension in the cohort according to birthweight and body the other five groups combined, 4.0% (95% CI: 3.54.5%,
mass index (BMI) (weight/height2) at age 11 years. We selected P for difference < 0.001). The corresponding figures for hyper-
11 years for comparability with previous studies,11 but the tension were 12.0% (95% CI: 10.213.9%) and 17.1% (95%
results for adjacent years are similar. In the past boys and girls CI: 16.018.1%, P , 0.001). The patterns of odds ratios and
of these ages in Finland would not yet have reached puberty. incidence shown in Tables 1 and 2 were similar in the two sexes.
The average age of menarche at that time was around 14 years.16
Coronary heart disease
We calculated the odds ratios for men and women combined
and adjusted them for year of birth and sex. Table 3 is based on 1235 patients who were admitted to hospital
We used the cumulative incidence of type 2 diabetes and or died from CHD, and 480 patients who died from the disease
hypertension to examine the proportion of cases attributable to
Table 1 Odds ratiosa (95% CI) for type 2 diabetes and hypertension
low birthweight and rapid childhood gain in BMI. We examined according to birthweight and body mass index at 11 years: 13 517 men
change in BMI between age 3 years and age 11 years because and women born 19241944
type 2 diabetes is associated with low rates of growth up to 2
age 2 years. Age 311 years broadly corresponds to the juvenile Birthweight Body mass index at 11 years (kg/m )
(kg) 15.7 16.6 17.6 .17.6
phase of growth.16 Use of age 2 or 4 years as the starting point
has little effect on the results. The analysis of cumulative No. of men and women
incidence was restricted to the people who were born 3.0 991 719 581 560
19341944, for whom growth data were available from birth, 3.5 1394 1422 1264 1246
through infancy and early childhood and into school years. As 4.0 827 984 1122 1110
described previously,912 we used standard deviation scores to .4.0 167 254 413 463
describe growth and divided the subjects into those whose Type 2 diabetes (698 cases)
standard deviation score for BMI increased or decreased 3.0 1.3 (0.62.8) 1.3 (0.62.8) 1.5 (0.73.4) 2.5 (1.25.5)
between the ages of 3 and 11 years. 3.5 1.0 (0.52.1) 1.0 (0.52.1) 1.5 (0.73.2) 1.7 (0.83.5)
To correspond with our analyses of type 2 diabetes and 4.0 1.0 (0.52.2) 0.9 (0.41.9) 0.9 (0.42.0) 1.7 (0.83.6)
hypertension we calculated hazard ratios, adjusted for sex and .4.0 1.0 1.1 (0.42.7) 0.7 (0.31.7) 1.2 (0.52.7)
year of birth, for death and for hospital discharge or death from
Hypertension (2997 cases)
CHD. We examined the associations with birthweight and BMI
3.0 2.0 (1.33.2) 1.9 (1.23.1) 1.9 (1.23.0) 2.3 (1.53.8)
at age 11 years in the total cohort of 13 517 men and women. We
3.5 1.7 (1.12.6) 1.9 (1.22.9) 1.9 (1.23.0) 2.2 (1.43.4)
again used cumulative incidence to examine the associations
4.0 1.7 (1.02.6) 1.7 (1.12.6) 1.5 (1.02.4) 1.9 (1.22.9)
with birth size and change in BMI between age 3 and 11 years
in younger subjects, born 19341944. Among men in the total .4.0 1.0 1.9 (1.13.1) 1.0 (0.61.7) 1.7 (1.12.8)
cohort, ponderal index at birth (birthweight/birth length3) a Odds ratios adjusted for sex and year of birth.
FETAL ORIGINS OF ADULT DISEASE 1237

Table 2 Cumulative incidence (%) of type 2 diabetes and Table 4 Cumulative incidence (%) of coronary heart disease (hospital
hypertension according to birthweight and change in standard admissions and deaths) according to body size at birth and change in
deviation score for body mass index between 3 and 11 years of age: standard deviation score for body mass index between 3 and 11 years:
6424 men and women born 19341944 3387 men and 2958 women born 19341944

Change in standard deviation Change in standard deviation


score for body mass index score for body mass index
311 years 311 years
Birthweight (kg) Decrease Increase Birth size Decrease Increase
Type 2 diabetes (227 cases) Men (279 cases)
3.2 3.1 (1075) 5.5 (1080) Birthweight (kg)
3.6 2.4 (1274) 4.3 (950) 3.2 8.8 (512) 9.0 (476)
.3.6 1.5 (1190) 5.4 (855) 3.6 6.9 (662) 11.3 (512)
Hypertension (1036 cases) .3.6 5.9 (740) 8.6 (521)
3.2 15.9 (1075) 21.3 (1080) Birth length (cm)
3.6 14.8 (1274) 19.4 (950) 49 8.1 (431) 9.6 (353)
.3.6 12.0 (1190) 13.9 (855) 50 6.9 (433) 9.0 (344)
Figures in parentheses are numbers of subjects. .50 6.7 (1034) 10.1 (792)
Ponderal index (kg/m3)
25 8.0 (411) 11.7 (394)
Table 3 Hazard ratiosa
(95% CI) for coronary heart disease according 27 7.6 (649) 10.8 (556)
to birthweight and body mass index at 11 years: 13 517 men and .27 6.2 (838) 7.2 (539)
women born 19241944
Women (66 cases)
2
Birthweight Body mass index at 11 years (kg/m ) Birthweight (kg)
(kg) 15.7 16.6 17.6 .17.6 3.2 1.6 (563) 3.8 (604)
Hospital admissions and deaths (1235 cases) 3.6 1.5 (612) 2.5 (438)
3.0 1.4 (0.82.4) 1.6 (0.92.8) 1.8 (1.03.2) 2.1 (1.13.8) .3.6 0.7 (450) 3.6 (334)
3.5 1.3 (0.72.2) 1.5 (0.92.7) 1.5 (0.82.6) 1.6 (0.92.9) Birth length (cm)
4.0 1.3 (0.72.3) 1.4 (0.82.4) 1.3 (0.82.4) 1.4 (0.82.6) 49 1.5 (543) 4.2 (520)
.4.0 1.0 1.2 (0.62.3) 1.1 (0.62.1) 1.0 (0.51.8) 50 1.5 (452) 3.3 (338)
Deaths (480 cases) .50 0.8 (609) 2.6 (496)
3.0 1.4 (0.54.0) 1.8 (0.65.1) 2.1 (0.76.2) 3.0 (1.08.6) Ponderal index (kg/m3)
3.5 1.4 (0.53.9) 1.9 (0.75.2) 2.2 (0.86.1) 2.7 (1.07.6) 25 1.6 (313) 4.3 (350)
4.0 1.9 (0.75.3) 1.8 (0.75.2) 1.7 (0.64.8) 1.6 (0.64.5) 27 1.5 (543) 2.4 (467)
.4.0 1.0 1.4 (0.44.6) 1.6 (0.54.7) 1.3 (0.44.0) .27 0.9 (748) 3.7 (537)
a Hazard ratios adjusted for sex and year of birth. Figures in parentheses are numbers of subjects. Seventy-nine subjects had
unknown birth length.

in people born 19241944. It shows hazard ratios according to length and change in BMI score among women were 0.8%
birthweight and fourths of BMI at age 11 years. Similarly to (95% CI: 0.31.9%) and 2.6% (95% CI: 2.03.3%, P = 0.02).
type 2 diabetes and hypertension (Table 1), the risks of disease Among men, CHD is related to low weight gain in infancy
fell with increasing birthweight and rose with increasing BMI. and to low BMI at age 1 year more strongly than to ponderal
The pattern was similar in the two sexes. The hazard ratios for index at birth.1,9 Table 5 shows the independent effects of BMI
admissions and deaths were 0.80 (95% CI: 0.720.90) for each at age one year and change in BMI between age 3 and 11 years
kilogram increase in birthweight and 1.06 (95% CI: 1.031.10) in men and women. In men with the highest BMI at age one
for each kg/m2 increase in BMI at age 11 years. The hazard year and a subsequent decrease in standard deviation score
ratios for deaths alone were 0.83 (95% CI: 0.690.99) and 1.10 for BMI the cumulative incidence of CHD was 5.4% (95%
(95% CI: 1.041.16). CI: 4.07.2%) compared with 9.2% (95% CI: 8.110.3%,
Table 4 is based on 279 hospital admissions or deaths from P = 0.001) in the other five groups. The corresponding figures
CHD among men and 66 among women in people born 1934 for women were 1.2 (95% CI: 0.52.5) and 2.5 (95% CI: 1.93.2,
1944. Similarly to type 2 diabetes and hypertension (Table 2), P = 0.10).
there were independent effects of birth size and change in BMI
score. Among men ponderal index at birth was more strongly
Discussion
related to CHD than birthweight: among women length at birth
was stronger. Among men with the highest ponderal index at Size of effects
birth and a subsequent decrease in standard deviation score for Men and women who had birthweights above 4 kg and whose
BMI the cumulative incidence of CHD was 6.2% (95% CI: pre-pubertal BMI was in the lowest fourth had around half the
4.78.1%) compared with 8.9% (95% CI: 7.810.1%, P = 0.03) risk of type 2 diabetes and hypertension when compared with
in the other five groups. The corresponding figures for birth people who had birthweights below 3 kg but whose BMI was in
1238 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Table 5 Cumulative incidence (%) of coronary heart disease (hospital produce lasting changes in the offsprings physiology and
admissions and deaths) among men and women according to body metabolism.20,21 That this should be so is unsurprising. So-called
mass index at one year and change in standard deviation score for
phenotypic plasticity enables one genotype to give rise to
body mass index between 3 and 11 years: 3387 men and 2958 women
born 19341944 a range of different physiological or morphological states in
response to different environmental conditions during
Change in standard deviation development.22,23 Such gene-environment interactions are
score for body mass index
311 years ubiquitous in the development of all living things. The benefit
Body mass index at one year
(kg/m2) Decrease Increase
of developmental plasticity is that, in a changing environment,
it enables the production of phenotypes that are better matched
Men (279 cases)
to their environment than would be possible by the production
17 10.4 (424) 11.2 (716)
of the same phenotype in all environments.23
18.3 7.0 (642) 8.0 (461)
The different size of newborn human babies exemplifies
.18.3 5.4 (832) 9.0 (312) plasticity. The growth of babies has to be constrained by the size
Women (66 cases) of the mother, otherwise normal birth could not occur. Small
17 1.5 (482) 3.8 (819) women have small babies:24 in pregnancies after ovum
18.3 1.1 (552) 3.3 (335) donation they have small babies even if the woman donating
.18.3 1.2 (570) 2.0 (200) the egg is large.25 Babies may be small because their growth is
Figures in parentheses are numbers of subjects. constrained in this way or because they lack the nutrients for
growth. As McCance wrote long ago, The size attained in utero
depends on the services which the mother is able to supply.
the highest fourth (Table 1). Disease risk was related to the rate These are mainly food and accommodation.26 Since mothers
of increase in childhood BMI as well as to the BMI attained height or pelvic dimensions are generally not found to be
at any particular age.11,12 If each individual in the cohort important predictors of chronic disease in later life,27,28
had been in the highest third of birthweight and had lowered research into the fetal origins of disease has focussed on the
their standard deviation score for BMI between age 3 and 11 nutrient supply to the baby. The central concept is that despite
years (Table 2) the incidence of type 2 diabetes would have been current levels of nutrition in Western countries the nutrition of
reduced by 57% and the incidence of hypertension by 25%. many fetuses and infants remains sub-optimal because the
There were similarly strong gradients in risk of CHD, whether nutrients available are unbalanced or because their delivery is
defined by death or by hospital admission or death, across constrained by the long and vulnerable fetal supply line.29 Size
groupings of birthweight and BMI at age 11 years (Table 3). at birth for gestational age serves as a marker of fetal
Again men and women who had birthweights above 4 kg nutrition.30
and whose pre-pubertal BMI was in the lowest fourth had When undernutrition during early development is followed
around half the risk of CHD when compared with people who by improved nutrition many animals and plants stage
had birthweights below 3 kg but whose BMI was in the highest accelerated or compensatory growth.31 Compensatory growth
fourth (Table 3). If each man in the cohort had been in the has costs, however, which in animals include reduced life-
highest third of ponderal index at birth, and each woman in span.31 One suggestion is that a higher rate of cell division
the highest third of birth length, and if each man or woman causes more rapid shortening of the protective ends of the
had lowered their BMI score between age 3 and 11 years the chromosomes (telomeres) and hastens cell death and organ
incidence of CHD would have been reduced by 25% in men and degradation.32 There are a number of other possible processes
63% in women (Table 4). The number of cases among women by which, in humans, undernutrition and small size at birth
is, however, small. Consistent with findings in Hertfordshire1 followed by rapid childhood growth could lead to cardio-
low weight and BMI at age 1 year were more strongly vascular disease and type 2 diabetes in later life.9,11,12 The
associated with CHD among men than low birthweight or Helsinki cohort has shown that people who were small at birth
ponderal index at birth.9 If each man had been in the highest and had rapid growth thereafter are not only physiologically
third of BMI at age 1 year, and had lowered his standard and morphologically different, but also have altered responses
deviation score for BMI between age 3 and 11 years the to poor living standards in later life.33
incidence of CHD would have been reduced by 40% (Table 5).
The large reductions in disease incidence that we have shown
to be associated with modest alterations in fetal, infant and child- Conclusion
hood growth may be underestimates. As Robinson17 wrote,
Our findings suggest that a substantial part of the risk of CHD,
birthweight and ponderal index (as well as body mass index)
type 2 diabetes and hypertension is established during
are crude measures of how fetal nutrition has affected body
development. We suggest that this reflects two widespread
composition, so the true size of the effect of fetal growth on
biological phenomena, developmental plasticity and com-
later disease is hard to measure.
pensatory growth.

Biological basis
Since McCance and Widdowsons pioneering observations in Acknowledgements
Cambridge,18,19 experimentalists have repeatedly demonstrated Funding: British Heart Foundation, Jahnsson Foundation,
that minor alterations to the diets of pregnant animals can Finska Lkaresllskapet.
FETAL ORIGINS OF ADULT DISEASE 1239

KEY MESSAGES
To quantify the importance of developmental processes in the genesis of chronic disease it is necessary to examine
the impact of fetal, infant and childhood growth on later death, hospital admission and treatment.
This has been examined in the Helsinki cohort.
Small body size at birth and during infancy followed by accelerated weight gain in childhood have large effects
on the incidence of coronary heart disease, type 2 diabetes and hypertension.
These diseases may originate through two biological phenomena, developmental plasticity and compensatory
growth.

References 15 Kangas T. The FinnDiab Report: Health Care of People with Diabetes in
1 Barker DJP, Osmond C, Winter PD, Margetts B, Simmonds SJ. Weight Finland: Inpatient and Outpatient Care, Metabolic Control and Direct Costs
in infancy and death from ischaemic heart disease. Lancet 1989;2: of Care. Helsinki: National Research and Development Centre for
57780. Welfare and Health, 1995. Report no. 58.
2 Frankel S, Elwood P, Sweetnam P, Yarnell J, Davey Smith G. 16 Tanner JM. Foetus into Man. Ware: Castlemead Publications, 1989.

Birthweight, body-mass index in middle age, and incident coronary 17 Robinson R. The fetal origins of adult disease. BMJ 2001;322:37576.
heart disease. Lancet 1996;348:147880. 18 McCance RA, Widdowson EM. Nutrition and growth. Proc R Soc Lond B
3 Rich-Edwards JW, Stampfer MJ, Manson JE et al. Birth weight and
1962;156:32637.
risk of cardiovascular disease in a cohort of women followed up since 19 McCance RA, Widdowson EM. The determinants of growth and form.
1976. BMJ 1997;315:396400.
4 Leon DA, Lithell HO, Vagero D et al. Reduced fetal growth rate and Proc R Soc Lond B 1974;185:117.
20 Desai M, Hales CN. Role of fetal and infant growth in programming
increased risk of death from ischaemic heart disease: cohort study of
15 000 Swedish men and women born 191529. BMJ 1998; metabolism in later life. Biol Rev 1997;72:32948.
21 Kwong WY, Wild A, Roberts P, Willis AC, Fleming TP. Maternal
317:24145.
5 Mi J, Law CM, Zhang KL, Osmond C, Stein CE, Barker DJP. Effects of undernutrition during the preimplantation period of rat development
infant birthweight and maternal body mass index in pregnancy on causes blastocyst abnormalities and programming of postnatal
components of the insulin resistance syndrome in China. Ann Intern hypertension. Development 2000;127:4195202.
Med 2000;132:25360. 22 Bateson P, Martin P. Design for a Life: How Behaviour Develops. London:
6 Lithell HO, McKeigue PM, Berglund L, Mohsen R, Lithell UB, Leon Jonathan Cape, 1999.
DA. Relation of size at birth to non-insulin dependent diabetes and 23 West-Eberhard MJ. Phenotypic plasticity and the origins of diversity.
insulin concentrations in men aged 5060 years. BMJ 1996; Ann Rev Ecol Syst 1989;20:24978.
312:40610. 24 Snow MHL, Sharp F, Fraser RB, Milner RDG (eds). Fetal Growth:
7 Huxley RR, Shiell AW, Law CM. The role of size at birth and postnatal
Proceedings of the 20th Study Group. London: Royal College of
catch-up growth in determining systolic blood pressure: a systematic Obstetricians and Gynaecologists, 1989. Effects of genome on fetal
review of the literature. J Hyptens 2000;18:81531. size at birth, pp. 111.
8 Brenner BM, Chertow GM. Congenital oligonephropathy and the 25 Brooks AA, Johnson MR, Steer PJ, Pawson ME, Abdalla HI. Birth
etiology of adult hypertension and progressive renal injury. Am J weight: nature or nurture? Early Hum Dev 1995;42:2935.
Kidney Dis 1994;23:17175. 26 McCance RA. Food, growth and time. Lancet 1962;2:62126.
9 Eriksson JG, Forsn T, Tuomilehto J, Osmond C, Barker DJP. Early
27 Barker DJP, Osmond C, Simmonds SJ, Wield GA. The relation of small
growth and coronary heart disease in later life: longitudinal study.
head circumference and thinness at birth to death from cardiovascular
BMJ 2001;322:94953.
disease in adult life. BMJ 1993;306:42226.
10 Forsn T, Eriksson JG, Tuomilehto J, Osmond C, Barker DJP. Growth
28 Forsen T, Eriksson JG, Tuomilehto J, Teramo K, Osmond C, Barker
in utero and during childhood among women who develop coronary
DJP. Mothers weight in pregnancy and coronary heart disease in a
heart disease: longitudinal study. BMJ 1999;319:140307.
cohort of Finnish men: follow up study. BMJ 1997;315:83740.
11 Forsn T, Eriksson J, Tuomilehto J, Reunanen A, Osmond C, Barker
29 Barker DJP. Mothers, Babies and Health in Later Life. 2nd Edn. Edinburgh:
DJP. The fetal and childhood growth of persons who develop type 2
Churchill Livingstone 1998.
diabetes. Ann Intern Med 2000;133:17682.
30 Harding JE. The nutritional basis of the fetal origins of adult disease.
12 Eriksson JG, Forsn T, Tuomilehto J, Osmond C, Barker DJP. Fetal and
childhood growth and hypertension in adult life. Hypertension Int J Epidemiol 2001;30:1523.
31 Metcalfe NB, Monaghan P. Compensation for a bad start: grow now,
2000;36:79094.
13 Hales CN, Barker DJP, Clark PMS et al. Fetal and infant growth and pay later. Trends Ecol Evol 2001;16:25460.
32 Hales CN. Suicide of the nephron. Lancet 2001;357:13637.
impaired glucose tolerance at age 64. BMJ 1991;303:101922.
14 WHO Study Group on Diabetes Mellitus. Diabetes Mellitus: Report of a 33 Barker DJP, Forsn T, Uutela A, Osmond C, Eriksson JG. Size at birth
WHO Study Group. Albany, NY: World Health Organization Centre and resilience to effects of poor living conditions in adult life:
USA, 1985. Technical Report Series no. 727. longitudinal study. BMJ 2001;323:127376.